research article the influence of aggregate size and...

Research ArticleThe Influence of Aggregate Size and Binder Material onthe Properties of Pervious Concrete

Tun Chi Fu, Weichung Yeih, Jiang Jhy Chang, and Ran Huang

Department of Harbor and River Engineering, National Taiwan Ocean University, 2 Pei-Ning Road, Keelung 20201, Taiwan

Correspondence should be addressed to Weichung Yeih; [email protected]

Received 21 August 2014; Revised 26 September 2014; Accepted 26 September 2014; Published 21 October 2014

Academic Editor: Aiguo Xu

Copyright © 2014 Tun Chi Fu et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Specimens were prepared by altering parameters such as aggregate sizes, bindermaterials, and the amounts of binder used and weresubsequently tested by using permeability, porosity, mechanical strength, and soundness tests. The results indicated that the waterpermeability coefficient and connected porosity decreased as the amount of binder used increased and increased with increasingaggregate size. In the mechanical strength test, the compressive, splitting tensile, and flexural strengths increased as the amountof binder used increased and decreased with the increase of aggregate size. Highly viscous binder enhanced compressive strength,water permeability, and the resistance to sulfate attacks. In the mechanics and sulfate soundness tests, the mix proportion of alkali-activated slag paste used in this study exhibited a superior performance than the Portland cement pervious concrete (the control)did, but the difference in water permeability between the two types of concrete was insignificant. The mix proportions of cementpaste containing 20% and 30% silica fume exhibited less mechanical strength than the control did. Moreover, compared with thecontrol, the cement paste containing silica fume demonstrated poor resistance to sulfate attacks, and the difference in the waterpermeability between such specimen and the control was not noticeable.

1. Introduction

Pervious concrete, also known as no-fines concrete, is a typeof porous compositematerial that can be regarded as concretecomposed of minimal to fine uniformly graded aggregatesand a limited amount of cement paste. Because of its perviousproperties, the pervious concrete pavement provides a bettertemperature and humidity exchange between atmosphereand earth than the impervious pavement such that the so-called “heat-island” effect in the urban region can be reduced.In addition, pervious concrete is regarded as a remedy for theflood control due to its excellent water permeability. Also, itis known that the pervious concrete pavement reduces trafficnoise and has an excellent antiskid performance.

However, a lower mechanical strength of pervious con-crete constrains the application of pervious concrete. Dueto the porosity inside the pervious concrete, its compressivestrength does not satisfy the minimal requirement for thestructural concrete (21MPa for 28-day compressive strength).Therefore, most applications for pervious concrete are park-ing lot pavement, pedestrian walkway, bike route, and places

where concrete compressive strength is not important. In thefollowing, literature survey for pervious concrete is given.

Generally, because of the high amount of connected voidsin pervious concrete, compressive strength is relatively low;thus, the strength of pervious concrete can be improved usingthe following strategies [1]:

(1) enhancing the characteristics of binder by decreasingthe water-cement (w/c) ratio and adding pozzolanicmaterials such as silica fume;

(2) adopting different binder materials such as usingepoxy to replace cement pastes;

(3) applying slight pressure and increasing the tempera-ture in the curing stage.

Marolf et al. studied the effect of aggregate size andgradation on the acoustic absorption for pervious concrete.They reported that pervious concrete mixtures with single-sized aggregates provide substantial improvement to soundabsorption as compared with conventional concrete [2]. Parket al. studied the sound absorption properties of pervious

Hindawi Publishing CorporationAdvances in Materials Science and EngineeringVolume 2014, Article ID 963971, 17 pageshttp://dx.doi.org/10.1155/2014/963971

2 Advances in Materials Science and Engineering

concrete using recycled aggregate and various target voidratios. They reported that the sound absorption charac-teristics of the porous concrete using recycled waste con-crete aggregate showed that the Noise Reduction Coefficient(NRC) was optimum at the void ratio of 25% but the percentcontent of the recycled aggregate had very little influence onthe NRC. Therefore, they concluded that the optimum voidratio is 25% and the recycled aggregate is 50% [3].

Park and Tia studied water purification effect of perviousconcrete. They found that a porous concrete with a smallersize of aggregate and a higher void content was found tohave superior ability of the removal of the total phosphorusand total nitrogen in the test water. They concluded thatthis effect is due to the large specific surface area of theporous concrete [4]. Huang et al. reported the properties ofpolymer-modified pervious concrete and they concluded thatit was possible to produce pervious concrete mixture withacceptable permeability and strength through the combina-tion of latex and sand [5]. Crouch et al. studied the aggregateeffect on the static elastic moduli of pervious concrete andthey reported that an increased aggregate amount resultedin a statistically significant decrease in both compressivestrength and static elastic moduli due to the subsequentdecrease in paste amount [6]. Chindaprasirt et al. studiedthe effects of aggregate size and binder strength on thecompressive strength and void ratio of pervious concrete [7].Chindaprasirt et al. also reported that good porous concreteswith void ratio of 15–25% and strength of 22–39MPa areproduced using paste with flow of 150–230mm and topsurface vibration of 10 s with vibrating energy of 90 kNm/m2[8].

Neithalath et al. used the values of porosity and themorphologically determined pore sizes, along with the porephase connectivity represented using an electrical conductiv-ity ratio, in a Katz-Thompson type relationship to predict thepermeability of pervious concretes [9]. Lian et al. developeda new model, which was based on the Griffth theory,to predict the compressive strength of pervious concreteusing its porosity [10]. Bentz used computer to simulatevarious virtual pervious concrete microstructural modelsand compares their percolation characteristics and computedtransport properties to those of real world pervious concretes[11].

Putman and Neptune evaluated different pervious con-crete test specimen preparation techniques in an effort toproduce specimens having properties similar to in-placepervious concrete pavement [12]. Lian and Zhuge studiedthe optimal mix design for pervious concrete [13]. Kim andLee studied the influence of cement flow and aggregate typeon mechanical and acoustic properties of pervious concrete[14]. Fujiwara et al. reported that a high strength perviousconcrete could be made by coating the coarse aggregateswith a high-strength mortar, then applying vibration to fusethem [15]. Safiuddin and Hearn compared the permeableporosity obtained from three different ASTM saturationtechniques, namely, cold-water saturation (CWS), boiling-water saturation (BWS), and vacuum saturation (VAS). Theyconcluded that vacuum saturation technique is more efficient

than cold-water or boiling-water saturation and thereforethis technique should be recommended for measuring thepermeable porosity of concrete [16]. Haselbach et al. devel-oped a theoretical model between the effective permeabilityof a sand-clogged pervious concrete block, the permeabilityof sand, and the porosity of the unclogged block [17]. Tho-In et al. have tried to use alkali-activated high-calcium flyash to make pervious concrete. They found that the high-calcium fly ash geopolymer binder could be used to producepervious concrete with satisfactory mechanical properties[18].

In this study, the performance of pervious concrete wasinvestigated by conducting various tests (i.e., mechanical,permeability, soundness, porosity, and unit weight tests)by changing parameters, which include aggregate size andthe type, amount, and w/c ratio of binders (for filling thevoids among aggregates). However, in addition to featuringexcellent water permeability, pervious concrete is expected toexhibit a mechanical performance to a certain level. Gener-ally, once pervious concrete has excellent water permeabilitythen its strength becomes poor. Various strategies can be usedto improve the strength of pervious concrete while achievingthe required permeability. By altering the parameters ofpervious concrete, this study obtained the optimal relation-ship between strength and water permeability. Subsequently,the required mix proportion was attained through relevantanalyses, and different binders were applied to multipletypes of pervious concrete to determine the mechanicalcharacteristics and water permeability of pervious concrete.In the experiment, the control group comprised pure cementpaste as bindingmaterial, and the experiment group involvedtwo types of cement pastes, one containing silica fume andone with alkali-activated slag. The feasibility of enhancingmechanical strength or changes in water permeability can beinvestigated using different binders.

2. Experiment Design

The experiment involved two stages. The first part involvedconsidering three crucial parameters, which were aggregatesize, the volume percent of aggregate voids filledwith binders,and the w/c ratio of binders, for determining the relationshipbetween water permeability and mechanical strength. Next,the feasibility of enhancing mechanical strength was deter-mined by varying the types of binders used based on the mixproportions obtained from the first part of the experiment.The second stage was aimed at investigating cement pastescontaining silica fume and alkali-activated slag. Varyingamounts of silica fume were used to replace cement; in otherwords, cement combined with pozzolanic material was usedinstead of pure cement paste. Alkaline-activated slag pastewith different liquid/slag (L/Sg) ratios was used to replace thepure cement pastes. Finally, the experiment and control (purecement paste only) groups were compared to investigate theinfluences of different binding materials on the properties ofpervious cement.

Pervious concrete is primarily composed of aggregatesand binding materials, which bind aggregates and fill voidsamong aggregates to form porous, permeable concrete.

Advances in Materials Science and Engineering 3

Table 1: Experimental variables of the two-stage experiment.

(a)

Experimental variables of Stage 1

Aggregate code and size A0.32 cm

B0.48 cm

C0.64 cm

D0.95 cm

Binder Cement pastew/c ratio 0.25 0.35 0.452Total volume percent of filled voids 50% 60% 70% 80%Aggregate code and size B (0.48 cm)

(b)

Experimental variables of Stage 2Binder Alkali-activated slag paste Silica-fume cement pasteTotal volume percent of filled voids 80%L/Sg or w/c ratio 0.35 0.4 0.45 0.35Substitution ratio — 10% 20% 30%

Regarding applications in engineering, pervious concreteshould have a certain level of water permeability before itsmechanical strength is enhanced. Therefore, designing thevoid content inside pervious concrete is the key that affectsthe overall property of thematerial. High porosity in perviousconcrete indicates excellent water permeability but poormechanical strength because of insufficient compactness. Bycontrast, low porosity can enhance mechanical strength butmight decrease water permeability. Although such judgmentcan be preliminarily made based on physics concepts, whenvariables such as binder characteristics and aggregate sizeare considered, whether or not the effect of porosity on theproperty of previous concrete conforms to common conceptsmust be determined by experimental studies. The parameterdesign concepts are explained as follows.

(i) Aggregate size: in this experiment, uniformly gradedaggregates (narrow gradation) were used, which areaggregates of a single size. In addition, aggregatesize affects the water permeability and mechanicalstrength of pervious concrete.

(ii) W/c ratio: in pervious concrete, when designingthe w/c ratios of cement pastes, workability, waterpermeability, andmechanical strength should be con-sidered. The workability of cement pastes influencesthe overall performance of pervious concrete.

(iii) The volume percent of aggregate voids filled withbinders (hereafter referred to as volume percent offilled voids): this volume percent refers to the amountof binders used in pervious concrete. This amountalso influences the water permeability and mechan-ical strength of pervious concrete.

(iv) Binders: mechanical strength is enhanced by alteringbinders, which influence the property of perviousconcrete.

The binders used in the first stage were cement pastesmade by mixing cement with water. Superplasticizers wereadded to the binder with a low w/c ratio of 0.25 to increase

the binder fluidity. In the second stage, cement pastescontaining silica fume and alkali-activated slag were used.Based on the first stage of the experiment, the control groupinvolved the following parameters: B aggregate size, w/c ratioof 0.35, and 80% volume percent of filled voids. Subsequently,comparison test was conducted in which the mix proportionof binders was altered. The experimental variables are shownin Table 1. In this table, the aggregate size means the mini-mum size of the sieve grid, for which aggregates pass throughthe sieve with greater sieve grid and stay in the current one.The total volume percent of filled voids can be understoodin the following. First, we packed aggregates inside a unitvolume andwe could calculate the volume of aggregate by theweight of aggregate and the specificweight of aggregate.Then,we subtracted the total volume by the aggregate volume andobtained the void volume. In such amanner, the total volumepercent of filled voids of 50% means that we use paste to fill50% void volume.

2.1. Materials. This study used the Portland Type I cementproduced and coarse aggregates of four different sizes as thosein Table 1. The aggregates were pebbles and their propertiesare presented in Table 2.

Superplasticizers are high-performance carboxylic plasti-cizers.The chemical compositions of sodium silicate, sodiumhydroxide, and phosphoric acid used for making alkali-activated slag paste are shown in Tables 3, 4, and 5. Thephosphoric acid is used as the retarder for avoid quick-settingproblem of the alkali-activated slag [19].

2.2. Mix Proportion Design

2.2.1. The Mix Proportion of Cement Paste as a Binder. Wefirst determined the void volume per unit volume for TypesA, B, C, and D aggregates as shown in Table 2. Subsequently,50%, 60%, 70%, and 80% void volumes were filled usingcement pastes. In addition, 5% superplasticizers (by thecement weight) were added to cement pastes with a w/c ratio


Table 2: Physical properties of coarse aggregates.

Code A B C DNominal size 1/8 in. 3/16 in. 1/4 in. 3/8 in.Maximum size 0.48 cm 0.64 cm 0.95 cm 1.27 cmRange of particle sizes 0.24–0.48 cm 0.48–0.64 cm 0.64–0.95 cm 0.95–1.27 cmSpecific gravity 2.65 2.69 2.66 2.72Void volume per unit volume 37.3% 37.5% 36.8% 38.3%

Table 3: Chemical composition of sodium silicate.

Item Sodium silicateTest item Test resultInsoluble residue Max. 0.01%Silicon dioxide (%) 37.0%Sodium oxide (%) 17.7%Mole ratio 2.16Iron (Fe) Max. 0.02%

Table 4: Chemical composition of sodium hydroxide.

Test item Test resultsChloride (Cl) Max. 0.005%Sulfate (SO4) Max. 0.003%Silicate (SiO2) Max. 0.01%Phosphate (PO4) Max. 0.001%Heavy metals (arsenic and lead) Max. 0.001%Iron (Fe) Max. 0.0007%Aluminum (Al) Max. 0.003%Calcium (Ca) Max. 0.001%Magnesium (Mg) Max. 0.0005%Potassium (K) Max. 0.1%Total nitrogen (N) Max. 0.001%Arsenic (As) Max. 0.0002%Sodium carbonate (Na2CO3) Max. 2.0%Assay (NaOH) Max. 95.0%

of 0.25 to enhance fluidity. The mix proportion design isshown in Table 6.

2.2.2. The Mix Proportion of Alkali-Activated Slag Paste as aBinder. The alkali activator was made by first adding sodiumhydroxide to sodium silicate and then mixing the mixtureuniformly before phosphoric acid (retarder) was added. Themix proportion design is shown in Table 7. In this table,the L/Sg ratio means the weight ratio between liquid phaseactivator and slag. The activator was made by mixing thehydroxide and sodium silicate and to reach the concentrationSiO2= 106 g/L and Na

2O = 105 g/L. And the retarder was

phosphoric acid of 0.74M.

2.2.3. The Mix Proportion of Silica-Fume Cement Paste as aBinder. Table 8 shows the mix proportion design for silica-fume cement paste; the substitution percentages for silicafume were 10%, 20%, and 30% of the cement weight.

Table 5: Chemical composition of phosphoric acid.

Item Phosphoric acidTest item Test resultsChloride (Cl) Max. 0.001%Nitrate (NO3) To pass testSulfate (SO4) Max. 0.006%Alkali and other phosphates (sulfate) To pass testSubstances reducing KMnO4 To pass testHeavy metals (As and Pb) Max. 0.001%Iron (Fe) Max. 0.005%Arsenic (As) Max. 0.0003%Specific gravity 1.700–1.710Assay Max. 85.0%

2.3. Tests

(1) Porosity test: the total porosity in pervious con-crete includes disconnected porosity and connectedporosity, which is the primary influencing factor ofwater permeability. A caliper was used to measureand calculate specimen volume 𝑉

1; the specimen was

immersed in water until it is filled with water beforeits weight in water𝑊

1is measured. Subsequently, the

specimen was taken out of water and dried, and thenits weight in air 𝑊

2when its weight is stable was

measured. The equation for connected porosity 𝑃1is

as follows:

𝑃1= [1 −(𝑊2−𝑊1)

𝑉1

] ∗ 100%. (1)

(2) Unit weight: after the pervious concrete specimenssolidified and were demolded, they were dried in anoven at 105 ± 5∘C until their weights were stable.The specimens were weighed and analyzed usinga caliper to measure, calculate, and obtain theirvolumes. Dividing the weight by volume yields theweight of pervious concrete per unit volume.

(3) Compressive strength: compressive strength wasdetermined based on the ASTM C39 for cylindricalconcrete specimens.

(4) Flexural strength: flexural strength was determinedbased on the three-point bending test.

(5) Splitting tensile strength: this test was performed fordetermining the splitting tensile strength of cylindri-cal concrete specimens.


Table 6: Testing proportions of cement pastes as a binder.

Binder Aggregate W/c ratio Percent ofvoids filled

Amount ofaggregate(kg/m3)

Amount ofcement (kg/m3)

Amount ofmixing water

(kg/m3)

Amount ofplasticizer(kg/m3)

Cement paste

A

0.25

50% 1665 328 66 1760% 1665 394 79 2070% 1665 460 92 2380% 1665 525 105 26

0.35

50% 1665 279 96 —60% 1665 335 115 —70% 1665 391 134 —80% 1665 447 154 —

0.45

50% 1665 243 109 —60% 1665 291 131 —70% 1665 340 153 —80% 1665 389 175 —

B

0.25

50% 1684 330 66 1760% 1684 396 79 2070% 1684 462 92 2380% 1684 568 106 26

0.35

50% 1684 281 98 —60% 1684 337 118 —70% 1684 393 138 —80% 1684 449 157 —

0.45

50% 1684 244 110 —60% 1684 293 132 —70% 1684 342 154 —80% 1684 391 176 —

C

0.25

50% 1683 324 65 1660% 1683 389 78 1970% 1683 454 91 2380% 1683 518 104 26

0.35

50% 1683 276 96 —60% 1683 331 116 —70% 1683 386 135 —80% 1683 441 154 —

0.45

50% 1683 240 108 —60% 1683 288 129 —70% 1683 336 151 —80% 1683 384 173 —

D

0.25

50% 1682 337 67 1760% 1682 404 81 2070% 1682 471 94 2480% 1682 530 108 27

0.35

50% 1682 286 100 —60% 1682 343 120 —70% 1682 401 140 —80% 1682 458 160 —


Table 6: Continued.

Binder Aggregate W/c ratio Percent ofvoids filled

Amount ofaggregate(kg/m3)

Amount ofcement (kg/m3)

Amount ofmixing water

(kg/m3)

Amount ofplasticizer(kg/m3)

0.45

50% 1682 249 112 —

60% 1682 299 134 —

70% 1682 349 157 —

80% 1682 398 179 —

Table 7: The mix proportion of alkali-activated slag paste.

L/Sg 0.35 0.40 0.45Aggregate size BVolume percent of filled voids 80%Concentration of alkali activator SiO2 = 106 g/L Na2O = 105 g/LAmount of phosphoric acid 0.74MAmount of slag (kg/m3) 432 403 377Amount of alkali-activated solution (kg/m3) 151 161 170Amount of aggregate (kg/m3) 1684 1684 1684

Table 8: The mix proportion of silica-fume cement paste.

Proportion of cement replaced by silica fume 10% 20% 30%Aggregate size BVolume percent of filled voids 80%Amount of cement (kg/m3) 404 359 314Amount of silica fume (kg/m3) 39 78 118Amount of mixing water (kg/m3) 155 153 151Amount of aggregate (kg/m3) 1684

(6) Water permeability coefficient: thewater permeabilitycoefficient was calculated using the constant-headpermeability test, which is based on the PavementTestManual established by the Japan Road Association.The permeability instrument measured the perme-ability coefficient of Φ10∗20 cm cylindrical speci-mens.The equation for water permeability coefficient𝐾 is expressed as follows:

𝐾 =𝑄𝐿

𝐴𝐻Δ𝑡, (2)

where𝐾 is water permeability coefficient (cm/s),𝑄 isflow volume (mL), 𝐿 is specimen thickness (cm), 𝐴is the pervious surface area of specimens (cm2),𝐻 iswater head height (cm), and Δ𝑡 = 𝑡

1− 𝑡0represents

time duration of measurement (s).(7) Soundness test: aggregate soundness tests using

magnesium sulfate were employed. Specimens wereplaced in an oven at 110 ± 5∘C to dry until theconstant weight 𝑊

1was obtained. After cooled to

room temperature, the specimen was fully immersed

(liquid rising 13mm above the top of the specimen)for at least 16 h but less than 18 h in an oversaturatedmagnesium sulfate solution. After immersion, thespecimen was taken out of the solution, washed usingclean water, and then dried until the constant weight𝑊2was obtained. The equation for calculating speci-

men weight-loss percentage is expressed as follows:

Weight loss% = 𝑤1 − 𝑤2𝑤1

× 100%. (3)

3. Test Results and Analysis

3.1. Unit Weight Test. Because pervious concrete has numer-ous voids, its unit weight is slightly lighter than those ofcommon concrete. Regarding the mix proportion of cementpastes, changes in unit weight were observed based on a fixedaggregate size. The unit weight increased with the amount ofbinders but decreasedwith the increase of w/c ratios as shownin Figures 1(a)–1(d). Comparing the influence of differentbinders on unit weight, the unit weight of alkali-activated slag


50 60 70 80

0

1000

2000

3000

Uni

t wei

ght (

kg/m

3)

Particle AFill percentage of voids by binder (%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a) Unit weights for particle A

50 60 70 80

0

1000

2000

3000

Uni

t wei

ght (

kg/m

3)

Fill percentage of voids by binder (%)Particle B

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b) Unit weights for particle B

Uni

t wei

ght (

kg/m

3)

Particle CFill percentage of voids by binder (%)50 60 70 80

0

1000

2000

3000

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c) Unit weights for particle C

50 60 70 80

0

1000

2000

3000

Uni

t wei

ght (

kg/m

3)

Particle DFill percentage of voids by binder (%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d) Unit weights for particle D

Figure 1: Unit weight of specimens comprising various particle sizes.

paste was greater than that of the control specimen (w/c =0.35; aggregate code =B; volume percent of filled voids = 80%;binder = Portland cement paste), whereas the unit weightof the silica-fume cement paste was smaller than that ofcontrol specimen. In addition, the unit weights of specimensdecreased as the amount of silica fume increased.

3.2. Permeability Test. Table 9 shows the water permeabilitycoefficients of the specimens with different mix proportions.The specimen label is explained in the following.The first two

digits represent the w/c ratio (Portland cement paste) or L/Sgratio (alkali-activated slag paste) or w/b ratio (silica-fumemodified Portland cement paste). The alphabet followed thetwo digits means the aggregate code (A, B, C, and D). Thenext two digits mean the volume percent of filled voids.After these, the remaining characters mean the binder type(C stands for cement paste; A stands for the alkali-activatedslag paste; S1 stands for silica-fume modified cement pastewith the replacement percentage of 10%; S2 stands for silica-fumemodified cement paste with the replacement percentage


Table 9: The water permeability coefficients for various mix proportions.

Label 𝐾 (cm/sec) Label 𝐾 (cm/sec) Label 𝐾 (cm/sec) Label 𝐾 (cm/sec)25A50C 0.1140 35A50C 0.0997 45A50C 0.0881 35B80A 0.043025A60C 0.1052 35A60C 0.0555 45A60C 0.0464 40B80A 0.044625A70C 0.0852 35A70C 0.0323 45A70C NA 45B80A 0.042125A80C 0.0436 35A80C 0.0304 45A80C NA 35B80S1 0.045425B50C 0.1259 35B50C 0.1192 45B50C 0.1176 35B80S2 0.044125B60C 0.1163 35B60C 0.1044 45B60C 0.0925 35B80S3 0.044425B70C 0.0864 35B70C 0.0585 45B70C NA25B80C 0.0857 35B80C 0.0439 45B80C NA25C50C 0.1423 35C50C 0.1273 45C50C 0.121025C60C 0.1283 35C60C 0.1199 45C60C 0.109725C70C 0.1185 35C70C 0.1118 45C70C NA25C80C 0.1126 35C80C 0.1046 45C80C NA25D50C 0.1440 35D50C 0.1355 45D50C 0.133125D60C 0.1350 35D60C 0.1292 45D60C 0.128325D70C 0.1284 35D70C 0.1131 45D70C NA25D80C 0.1149 35D80C 0.1105 45D80C NA

of 20%; and S3 stands for silica-fume modified cementpaste with the replacement percentage of 30%). For example,the label 25A50C means w/c = 0.25; aggregate code =A; volume percentage of filled voids = 50%; and binder =cement paste. In Table 9, there exist some specimens withoutpermeability coefficients. The sagging phenomenon, whichmeans that the cement paste does not stick to the aggregateand sags to the bottom of specimen due to its gravity, wasobserved. Once the sagging phenomenon happens, the pastewill not be uniformly distributed around aggregates andwill be accumulated in the bottom of specimen due to thegravity. Consequently, the water permeability will dramati-cally decrease since on the cross-section near the bottom thepermeability is low. Based on this, we discard the recordsof these specimens. This result indicated that for w/c =0.45, the viscosity enhancer should be introduced to avoid thepossible sagging phenomenon.

The mix proportion of particle D exhibited an optimalwater permeability coefficient 𝐾. Specifically, the specimen25D50C with aggregate D, w/c ratio of 0.25, and 50% binder-filled voids had the greatest coefficient 𝐾 of 0.1440 cm/sec.However, the specimen 35A80C with particle A, w/c ratioof 0.35, and 80% binder-filled voids had the smallest coef-ficient 𝐾 of 0.0304. Figures 2(a)–2(d) show the graphscomparing the particle size variables based on data fromTable 5.

For the different aggregate sizes, the water permeabilitycoefficients decreased as binder w/c ratios increased. This isbecause at low w/c ratio, binders are highly viscous, enablingthe binder to fully cover the aggregates. Intact and fully cov-ered aggregate forms after specimens solidified, and point-to-point contact among the aggregates remains, yielding a voidstructure that forms a path for water permeation. By contrast,with high w/c ratios, because binders have great fluidity,

excess binders that cannot cover aggregates may block thepath, thereby decreasing water permeability.

In addition, as the aggregate size increased, the waterpermeability of pervious concrete increased. For the con-ventional concrete, three phases (mortar, interface transitionzone, and coarse aggregates) exist. The interface transitionzone (ITZ) has the worst property among these three phases.It means that the mechanical strength of ITZ is weakestand the water permeability of ITZ is highest since themicrostructure of ITZ contains moremicrocracks, voids, andso on. Therefore, it is well known that the ITZ dominatesthe behaviors of conventional concrete. If the volume fractionof ITZ increases, a lower compressive strength and a higherpermeability coefficient then are expected. Consequently, asthe aggregate size was smaller, the volume fraction of ITZwas expected to be larger. It then becomes very puzzlingwhy the water permeability coefficient increased as theaggregate size increased. The reason is explained as thefollowing. For the pervious concrete, four phases (paste,ITZ, coarse aggregates, and designed porosity) exist. Unlikethe conventional concrete which adopts mortar to fill thespace between aggregates, we only partially fill the spacebetween aggregates by binder for pervious concrete. Thisconcept intentionally makes some porosity inside concrete inorder to allow water penetration. Among these four phases,the designed porosity has the worst properties. It has nomechanical strength at all and highest water permeationcoefficient. Therefore, the volume fraction of the porositydominates the behaviors of pervious concrete. When aggre-gate size is smaller, the initial porosity after packing is smaller.Consequently, when the volume percent of filled voids is thesame, the remaining porosity for pervious concrete madewith smaller size aggregates is smaller. Therefore, the waterpermeability coefficient becomes lower.


50 60 70 80

0

0.02

0.04

0.06

0.08

0.1

0.12K

(cm

/s)

Fill percentage of voids by binder (%)Particle A

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a)

50 60 70 80

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

K(c

m/s

)

Fill percentage of voids by binder (%)Particle B

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b)

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

50 60 70 80

Fill percentage of voids by binder (%)

K(c

m/s

)

Particle CBinder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c)

50 60 70 80

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16


K(c

m/s

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d)

Figure 2: Water permeability coefficients of specimens comprising different aggregate sizes: (a) aggregate A; (b) aggregate B; (c) aggregate C;(d) aggregate D.

3.3. Porosity Test. The internal structure of pervious con-crete is relatively less compact because using an insufficientamount of binders produces voids. Water permeability isprimarily affected by connected porosity, which can bemeasured through experimental tests. As shown in Figures3(a)–3(d), a fixed trend was observed when other parameterswere compared based on the various mix proportions. Theporosity in pervious concrete is typically inversely propor-tional to the amount of binders used. Basically, the trends for

the values of connected porosity were similar to those forwater permeability coefficient. It is worth mentioning herethat the connected porosity increased as the aggregate sizeincreased. The reason has been mentioned in the previoussubsection. Once again, the dominating factor is the totalvoids inside pervious concrete but not the ITZ.

3.4. Compressive Strength Test. Table 10 shows the resultsof 28-day compressive strength tests for various mix


Table 10: The results of the compressive strength test.

Label Compressivestrength (MPa) Labe1 Compressive

strength (MPa) Label Compressivestrength (MPa) Label Compressive

strength (MPa)25A50C 17.49 35A50C 12.30 45A50C 10.62 35B80A 28.7325A60C 19.01 35A60C 16.01 45A60C 14.83 40B80A 24.9825A70C 22.61 35A70C 20.95 45A70C — 45B80A 17.4925A80C 25.67 35A80C 24.94 45A80C — 35B80S1 21.9825B50C 17.07 35B50C 11.87 45B50C 8.74 35B80S2 15.4325B60C 17.45 35B60C 14.49 45B60C 11.62 35B80S3 7.6825B70C 20.36 35B70C 18.81 45B70C —25B80C 21.95 35B80C 20.86 45B80C —25C50C 16.26 35C50C 11.24 45C50C 7.4925C60C 16.86 35C60C 13.18 45C60C 9.4925C70C 20.01 35C70C 15.51 45C70C —25C80C 21.38 35C80C 16.49 45C80C —25D50C 13.45 35D50C 10.34 45D50C 5.6225D60C 13.68 35D60C 11.87 45D60C 7.9325D70C 17.32 35D70C 14.86 45D70C —25D80C 19.20 35D80C 15.61 45D80C —

proportions.The alkali-activated slag paste (35B80A) yieldedthe highest compressive strength of 28.73MPa, and thecement paste (45D50C) yielded the smallest strength of5.62MPa. After compiling the data obtained from Table 6,Figures 4(a)-4(b)were plotted, displaying a comparison of theparameters based on fixed particle sizes.

It can be found that the compressive strength decreasedas the aggregate size increased. Similar results were alsofound for othermechanical properties such as splitting tensilestrength test and flexural strength test. As explained above,for the conventional concrete, the property of ITZ dominatesthe property of concrete since ITZ plays as the weakestpart in concrete. However, for pervious concrete the weakestpart now is designed void. As an increase in aggregatesize, a higher volume of void then is expected under thesame condition. In such a manner, the compressive strengthdecreased. Similar result has been observed in [4].

3.5. Splitting Tensile Strength Test. Figures 5(a)–5(d) showthe results of the splitting tensile strength tests. The alkali-activated slag paste (35B80A) yielded the highest splittingtensile strength of 3.19MPa, and the cement paste (45D50C)yielded the smallest strength of 1.21MPa. According toFigures 5(a)–5(d), the binder with a w/c ratio of 0.25 had thegreatest splitting tensile strength, whereas the binder with aw/c ratio of 0.45 had the weakest splitting tensile strength.Therefore, the w/c ratios of binders influence splitting ten-sile strength. In addition, using a substantial amount ofbinder strengthened splitting tensile strength; moreover, thespecimens with 80% binder-filled voids yielded the highestsplitting tensile strength.

Furthermore, the splitting tensile strength decreased asthe aggregate size increased. The reason has been given inSection 3.4.

3.6. Flexural Strength Test. Figures 6(a)–6(d) present theresults of flexural strength tests. Specimen 35B80C exhibitedthe highest flexural strength at 4.76MPa, and specimen45D50C produced the weakest flexural strength of 2.18MPa.The figures show that flexural strength is directly propor-tional to the amount of binder used. In addition, the flexuralstrength of binders with low w/c ratio is superior to thatof high-w/c ratio binders. Thus, the mix proportions withhighest flexural strengths had a w/c ratio of 0.25 and 80%binder-filled voids, and those with weakest flexural strengthshad a w/c ratio of 0.45 and 50% binder-filled voids. Inaddition, the flexural strength decreased as the aggregate sizeincreased. The reason has been given in Section 3.4.

3.7. Soundness Test. Soundness tests can simulate the resis-tance of pervious concrete specimens to sulfate attacks.In this study, 12-day old specimens were immersed in anoversaturated magnesium sulfate solution and then washedbefore subjecting to a round robin test. Subsequently, theweight loss percentages of the specimens were measured toinfer the resistance to sulfate attacks; the results are shownin Figures 7(a)–7(d). Specimen 35B80S3, which comprisedcement paste and 30% silica fume, exhibited a great weightloss percentage of 3%, and the alkali-activated slag specimen35B80A yielded a minimal weight loss percentage of 0.57%.

For all the aggregate sizes,maximumweight loss occurredwith 50% binder-filled void, and weight loss increased asthe amount of binder used increased. Regarding bindercharacteristics, weight loss was minimal at a w/c ratio of0.25 and increased with increasing w/c ratio.This was causedby the overall compactness of the specimen. When a smallamount of binder was used, increased number of voids in thespecimen became available for sulfate attacks and reactionsbetween cements and aggregates. As with w/c ratios, after


50 60 70 80

0

4

8

12

16

20


Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a)

50 60 70 80

0

5

10

15

20

25

Particle BFill percentage of voids by binder (%)

Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b)

50 60 70 80

0

5

10

15

20

25

Particle CFill percentage of voids by binder (%)

Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c)

50 60 70 80

0

5

10

15

20

25


Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d)

Figure 3: Connected porosity of specimens comprising different aggregate sizes: (a) aggregateA; (b) aggregate B; (c) aggregateC; (d) aggregateD.

high-w/c binders solidified, the amount of small voids thatremained after water loss or hydration was greater than thatof low-w/c binders. This increase in void amount createdpaths for sulfate attacks, thus decreasing the sulfate-attackresistance of high-w/c pervious concrete containing lowamount of binders.

3.8. The Cross-Comparison of Water Permeability Coefficients.The specimens comprising four various aggregate sizes weresubjected to permeability and connected porosity tests. The

test results were then used to plot an 𝑋-𝑌 distribution graphto which trend lines were added as shown in Figure 8. Forthe four types of particle sizes, water permeability coefficient𝐾 was directly proportional to connected porosity. Next,an analysis on water permeability coefficients based on afixed porosity was conducted, which revealed a directlyproportional relationship between permeability and aggre-gate size. From a physics point of view, a fixed porosityshould only correspond with a specific permeability coeffi-cient, which contradicts the results of this study, in which


50 60 70 80

0

4

8

12

16

20


Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a)

50 60 70 80

0

5

10

15

20

25


Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b)

50 60 70 80

0

5

10

15

20

25


Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c)

50 60 70 80

0

5

10

15

20

25


Void

(%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d)

Figure 4: Compressive strengths of specimens comprising different aggregate sizes: (a) aggregate A; (b) aggregate B; (c) aggregate C; (d)aggregate D.

multiple coefficients were observed.The reasonmight be thatalthough the specimens had the same porosity under certainconditions, the void sizes differ depending on the aggregatesize. Under identical connected porosity, the specimens withsmall aggregates had small sectional areas in the connectedvoids, generating meandering paths for water permeation.Conversely, specimens consisting of large particles had largesectional areas and, thus, had straight paths. Therefore, the

velocity (𝑉) of the water infiltrated in the specimen differed,yielding variation in the flow volume (𝑄) of the water flowingout of the specimen.

To determine the correlation between compressivestrength and water permeability, these two variables weremeasured for each particle size and then an𝑋-𝑌 distributiongraph was plotted. Trend lines were added to the graph toillustrate the correlation. Figure 9 shows that mechanical


50 60 70 80

0

1

2

3

4

5


Split

ting

stren

gth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a)

50 60 70 80

0

1

2

3

4


Split

ting

stren

gth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b)

50 60 70 80

0

1

2

3

4


Split

ting

stren

gth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c)

50 60 70 80

0

1

2

3


Split

ting

stren

gth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d)

Figure 5: Splitting tensile strengths of specimens comprising different aggregate sizes: (a) aggregate A; (b) aggregate B; (c) aggregate C; (d)aggregate D.

strength decreases as water permeability coefficient increases,and such relationship is applicable for all aggregate sizes.

4. Conclusions

(1) Water permeability coefficient and connected poros-ity decreased with the increase of binder amounts butincreased with increasing aggregate size.

(2) Pervious concrete with binders of low w/c ratio ishighly viscous, which facilitated covering the aggre-gates. This enabled sufficient binding between par-ticles and effectively reduced excess binders fromblocking water permeation paths, thereby influencingwater permeability.

(3) Under identical connected porosity, the specimenswith small particles had small sectional areas inthe connected voids, producing meandering paths


50 60 70 80

0

3

4

5


Flex

ural

stre

ngth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a)

50 60 70 80

0

2

3

4

Fill percentage of voids by binder (%)

Flex

ural

stre

ngth

(MPa

)

Particle BBinder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b)

50 60 70 80

0

2

3

4


Flex

ural

stre

ngth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c)

50 60 70 80

0

2

3

4


Flex

ural

stre

ngth

(MPa

)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d)

Figure 6: Flexural strengths of specimens comprising different particle sizes: (a) aggregate A; (b) aggregate B; (c) aggregate C; (d) aggregateD.

for water permeation. Specimens consisting of largeparticles had large sectional areas and straight paths.Thus, the velocity of the water infiltrated in the speci-men differed, yielding variation in the flow volume ofthe water flowing out of the specimen.

(4) Mechanical strength decreased with increasing waterpermeability. Although using substantial amountof binders can enhance mechanical strength, per-meability might decrease because the volume per-cent of binder-filled voids increases. The amount of

binder used was directly proportional to mechani-cal strength, and increased aggregate size decreasedmechanical strength.

(5) Alkali-activated slag paste with L/Sg = 0.35 and 0.4had greater mechanical strength than the cement-paste control did. Consequently, using appropriateamount of alkali-activated slag as a binder can effec-tively enhance the mechanical strength of perviousconcrete.


50 60 70 80

0

0.4

0.8

1.2

1.6

BinderFill percentage of voids by binder (%)

Wei

ght l

oss (

%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(a)

50 60 70 80

0

0.4

0.8

1.2

1.6

2


Wei

ght l

oss (

%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(b)

50 60 70 80

0

0.5

1

1.5

2

2.5


Wei

ght l

oss (

%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(c)

50 60 70 80

0

1

2

3


Wei

ght l

oss (

%)

Binder w/c = 0.25

Binder w/c = 0.35

Binder w/c = 0.45

(d)

Figure 7:Theweight loss percentages of specimens comprising different aggregate sizes that were immersed in a sulfate solution: (a) aggregateA; (b) aggregate B; (c) aggregate C; (d) aggregate D.

(6) The mechanical strengths of the specimens withbinders containing 10% silica fume were superior tothose of the control specimen. This result was notobserved for specimens with binders containing 20%or 30% silica fume, possibly because excessive amountof silica fume was added.Therefore, adding a suitableamount of silica fume to binders can enhance theoverall mechanical strength of pervious concrete.

(7) The weight-loss percentage of pervious concreteexperiencing sulfate attacks increased as the w/c ratioof cement pastes increased. Weight loss percentagewas directly proportional to aggregate size, becausethe stacking of large aggregates generated large voids,which created paths for sulfate to infiltrate specimensand induce reactions. Specimens composed of alkali-activated slag binder exhibited a superior resistance to


0 5 10 15 20 25

0

0.04

0.08

0.12

0.16

K(c

m/s

)

Void (%)Particle AParticle BParticle CParticle D

Particle AParticle BParticle CParticle D

Figure 8:The relationship between porosity and water permeabilitycoefficient.

5 10 15 20 25 30

0

0.04

0.08

0.12

0.16

K(c

m/s

)

Compressive strength (MPa)Particle AParticle B

Particle CParticle D

Figure 9: Correlation between compressive strength and waterpermeability coefficient.

sulfate attacks than those consisting of cement pastesdid.

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper.

References

[1] J. Yang and G. Jiang, “Experimental study on properties ofpervious concrete pavement materials,” Cement and ConcreteResearch, vol. 33, no. 3, pp. 381–386, 2003.

[2] A.Marolf, N.Neithalath, E. Sell, K.Wegner, J.Weiss, and J. Olek,“Influence of aggregate size and gradation on acoustic absorp-tion of enhanced porosity concrete,” ACI Materials Journal, vol.101, no. 1, pp. 82–91, 2004.

[3] S. B. Park, D. S. Seo, and J. Lee, “Studies on the sound absorptioncharacteristics of porous concrete based on the content ofrecycled aggregate and target void ratio,” Cement and ConcreteResearch, vol. 35, no. 9, pp. 1846–1854, 2005.

[4] S.-B. Park and M. Tia, “An experimental study on the water-purification properties of porous concrete,” Cement and Con-crete Research, vol. 34, no. 2, pp. 177–184, 2004.

[5] B. Huang, H. Wu, X. Shu, and E. G. Burdette, “Laboratoryevaluation of permeability and strength of polymer-modifiedpervious concrete,”Construction and BuildingMaterials, vol. 24,no. 5, pp. 818–823, 2010.

[6] L. K. Crouch, J. Pitt, and R. Hewitt, “Aggregate effects onpervious portland cement concrete static modulus of elasticity,”Journal of Materials in Civil Engineering, vol. 19, no. 7, pp. 561–568, 2007.

[7] P. Chindaprasirt, S. Hatanaka, N. Mishima, Y. Yuasa, and T.Chareerat, “Effects of binder strength and aggregate size onthe compressive strength and void ratio of porous concrete,”International Journal of Minerals, Metallurgy andMaterials, vol.16, no. 6, pp. 714–719, 2009.

[8] P. Chindaprasirt, S. Hatanaka, T. Chareerat, N. Mishima, andY. Yuasa, “Cement paste characteristics and porous concreteproperties,” Construction and Building Materials, vol. 22, no. 5,pp. 894–901, 2008.

[9] N. Neithalath, M. S. Sumanasooriya, and O. Deo, “Characteriz-ing pore volume, sizes, and connectivity in pervious concretesfor permeability prediction,”Materials Characterization, vol. 61,no. 8, pp. 802–813, 2010.

[10] C. Lian, Y. Zhuge, and S. Beecham, “The relationship betweenporosity and strength for porous concrete,” Construction andBuilding Materials, vol. 25, no. 11, pp. 4294–4298, 2011.

[11] D. R. Bentz, “Virtual pervious concrete: microstructure, perco-lation, and permeability,” ACI Materials Journal, vol. 105, no. 3,pp. 297–301, 2008.

[12] B. J. Putman and A. I. Neptune, “Comparison of test specimenpreparation techniques for pervious concrete pavements,” Con-struction and Building Materials, vol. 25, no. 8, pp. 3480–3485,2011.

[13] C. Lian and Y. Zhuge, “Optimum mix design of enhanced per-meable concrete—an experimental investigation,” Constructionand Building Materials, vol. 24, no. 12, pp. 2664–2671, 2010.

[14] H. K. Kim and H. K. Lee, “Influence of cement flow andaggregate type on themechanical and acoustic characteristics ofporous concrete,” Applied Acoustics, vol. 71, no. 7, pp. 607–615,2010.

[15] H. Fujiwara, R. Tomita, T. Okamoto, A. Dozono, and A.Obatake, “Properties of high-strength porous concrete,” ACIStructural Journal, vol. SP-179, pp. 173–188, 1998.

[16] M. Safiuddin and N. Hearn, “Comparison of ASTM saturationtechniques for measuring the permeable porosity of concrete,”Cement and Concrete Research, vol. 35, no. 5, pp. 1008–1013,2005.

[17] L. M. Haselbach, S. Valavala, and F. Montes, “Permeability pre-dictions for sand-clogged Portland cement pervious concretepavement systems,” Journal of Environmental Management, vol.81, no. 1, pp. 42–49, 2006.


[18] T.Tho-In, V. Sata, P. Chindaprasirt, and C. Jaturapitakkul, “Per-vious high-calcium fly ash geopolymer concrete,” Constructionand Building Materials, vol. 30, pp. 366–371, 2012.

[19] J. J. Chang, W. Yeih, and C. C. Hung, “Effects of gypsum andphosphoric acid on the properties of sodium silicate-basedalkali-activated slag pastes,” Cement and Concrete Composites,vol. 27, no. 1, pp. 85–91, 2005.

Submit your manuscripts athttp://www.hindawi.com

ScientificaHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

CorrosionInternational Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Polymer ScienceInternational Journal of



CeramicsJournal of


CompositesJournal of

NanoparticlesJournal of



International Journal of

Biomaterials


NanoscienceJournal of

TextilesHindawi Publishing Corporation http://www.hindawi.com Volume 2014

Journal of

NanotechnologyHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Journal of

CrystallographyJournal of


The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014


CoatingsJournal of

Advances in

Materials Science and EngineeringHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Smart Materials Research



MetallurgyJournal of


BioMed Research International

MaterialsJournal of


Nano

materials


Journal ofNanomaterials

research article the influence of aggregate size and...

Documents